[Ml-stat-talks] Fwd: ORFE Colloquium: Daniela Witten, February 28th at 4:30pm, Sherrerd Hall 101
bee at princeton.edu
Thu Feb 23 09:38:47 EST 2017
Talk of interest.
=== ORFE Colloquium Announcement ===
DATE: Tuesday, February 28, 2017
LOCATION: Sherrerd Hall, room 101
SPEAKER: Daniela Witten, University of Washington
TITLE: Graphical Modeling of Temporal Data
ABSTRACT: In this talk, I will consider the problem of learning the
structure of a graphical model on the basis of high-dimensional data
measurements taken over time.
First, I'll consider the setting of continuous-valued observations that are
governed by an additive ordinary differential equation model. Existing
approaches for learning the structure of the graph smooth the data over
time, estimate the derivatives, and then perform a nonparametric regression
of the estimated derivatives onto the features. I will show that a
seemingly small modification of this approach leads to much better results.
Next, I'll consider the case of multivariate point process data. I will
present a non-parametric approach for learning the structure of the graph,
for which theoretical guarantees are available. Furthermore, under the
assumption that the point process is "mutually exciting", a much simpler
screening approach has the "sure screening" property: with high
probability, it will avoid any false positive edges.
This is joint work with Ali Shojaie (University of Washington) and Shizhe
Chen (Columbia University).
BIO: Daniela Witten's research involves the development of statistical
machine learning methods for high-dimensional data, with applications to
genomics and other fields. She is particularly interested in unsupervised
learning, with a focus on graphical modeling. Daniela is the recipient of a
number of honors, including a NDSEG Research Fellowship, an NIH Director's
Early Independence Award, a Sloan Research Fellowship, and an NSF CAREER
Award. Her work has been featured in the popular media: among other forums,
in Forbes Magazine (three times), Elle Magazine, on KUOW radio, and as a
PopTech Science Fellow. Daniela is a co-author (with Gareth James, Trevor
Hastie, and Rob Tibshirani) of the very popular textbook "Introduction to
Statistical Learning". She was a member of the Institute of Medicine
committee that released the report "Evolution of Translational Omics".
Daniela completed a BS in Math and Biology with Honors and Distinction at
Stanford University in 2005, and a PhD in Statistics at Stanford University
in 2010. Since 2014, Daniela is an associate professor in Statistics and
Biostatistics at University of Washington.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Ml-stat-talks